Acquired disorders of reading (acquired dyslexias) are nearly always seen in patients with aphasia I subsequent to left hemisphere stroke. Even when language functions recover sufficiently to enable the patient to return to work, an unrecovered problem with reading often interferes significantly with job performance. Patients who cannot return to work may be left with little to occupy their time; the ability to read for pleasure could make a significant difference in their quality of life. The purpose of this project is to develop, implement, and test a set of therapy programs for the treatment of acquired dyslexias, based upon a cognitive neuropsychological model of reading. A comprehensive and detailed battery of reading and reading-related tests will be used to determine the underlying impairment causing the reading deficit in each patient. This proposal focuses on three specific deficits: (l) impaired access to the visual word form from the visual modality (pure dyslexia); (2) impaired activation of orthographic representations (surface dyslexia); and (3) impaired phonological reading route (phonologic/deep dyslexia). A set of experimental treatment programs has been devised for each of these three reading deficits. These treatment programs derive in part from cognitive models of reading, in part from differences in general approach to treatment, and in part from the results of previous treatment studies. All treatment programs will utilize a single-subject multiple-baseline design, in which a group of letters, words, rules, etc. will be learned to a criterion of 90% correct before the next group is trained. Each treatment program will be replicated in several patients. A set of general external probe tests will be administered to all subjects before and after treatment. There are additional external probes for each of the three deficit types, consisting of a list of words targeted for improvement but never trained. Measures include both accuracy and speed of reading. Treatment programs will be evaluated for efficacy by examining the multiple baseline graphs, comparing accuracy and speed of reading the treatment-specific probe words pre- and post-treatment, and examining performance on the general external probes pre- and post-treatment. Certain circumscribed variables hypothesized to have predictive value for a specific deficit group will be examined. The results of this project will serve as a guide for the speech/language pathologist in choosing an efficacious treatment tailored to the specific needs of the individual dyslexic patient. In addition, the data may be used to improve models of normal reading, which may lead to more effective methods of teaching reading to both normal and developmentally dyslexic children.